IRIS Modulo 1 - El factor de demanda en la era de la IA. Principio fiscal, arquitectura de medición y horizonte de financiamiento[IRIS Module 1 - Demand Factor in the Age of AI: Fiscal Principle, Measurement Architecture, and Financing Horizon]

Access to the human market was never free. Its price, invisible, was embedded in the wage bill: in hiring and paying workers for private reasons, firms reproduced, as an automatic byproduct, the aggregate demand from which they benefited. Automation does not eliminate that price: it leaves it unpaid. This Discussion Paper develops the consequences of that observation. Under the industrial economy, the wage performed a dual function: it remunerated labor and distributed the purchasing power that allowed production to be realized as market demand. An AI-intensive economy can unbundle these functions, raising productive capacity while reducing the centrality of employment as a channel of distribution. If the wage bill falls relative to output, the resulting gap can be closed through only five channels: consumption out of capital income, net exports, household debt, public transfers, or the deflationary destruction of capacity. Only the fourth is compatible with stability. From this follows a Copernican turn: the human being does not disappear from the economic circuit but changes function. From central producer, the human being becomes the indispensable bearer of the demand factor, the capacity to absorb production through purchasing power, granting it market value, social legitimacy, and economic purpose. The IRIS principle states that AI's access to that demand cannot remain institutionally free of charge. Not to punish innovation, but to reconstitute the demand-funding flow that substitution interrupts. Three conceptual instruments follow. The MAT (Market Access Tax) captures the uncompensated incidence of AI on domestic production, scaling with substitution rather than with profitability and bounded by an innovation non-braking principle. The MAR (Market Access Right) extends that logic when market access crosses jurisdictions. The MAD (Market Access Dividend) returns demand capacity to the social body, not as charity, but as institutional compensation for a collective economic function. The paper then develops a transitional architecture for measuring the incidence of AI, from technical, economic, and declarative proxies toward the hypothesis of an AI-assisted enterprise interpretive agent. Finally, it identifies the structural limit of taxation: even a well-designed MAT-MAR-MAD system, reinforced by verifiable gains in AI-assisted public efficiency, might not suffice to sustain broad effective demand in a deeply automated economy. IRIS is therefore configured as a bridge architecture: fiscal in its first movement, social in its second, monetary-institutional in its third.